On 1/22/07, Jeff Gerritsen <..hidden..> wrote:
> I believe a healthy discussion on the future of LSMB is needed, although I'm
> concerned about two issues, one being discussed and one not being discussed!
>
> The concerns I have are about the one issue being discussed may degenerate
> into unproductive language and platform wars - although (I believe) a remote
> possibility).
Given past conflicts, what will happen is that any language/platform
wars will cause an issue to be tabled and will eventually be revisited
and resolved. So far, the core team has been very good at working to
resolve these differences, though I am sure that both Josh and I are
surprised that the differences exist :-)
> The second issue not being discussed (at least I haven't seen
> much written about it) is what are our intended users needs? Shouldn't these
> needs be the driving force behind the future direction of LSMB? Therefore,
> based upon user needs, can we not make intelligent decisions on the future of
> LSMB?
You are right about this. I would just point out three things:
1) Most of the actual mechanics of how the application needs to work
are well defined by standards boards. I think that it is important
that LSMB always operates correctly from an accounting perspective.
Given the state of the inherited codebase, it will be a while before
this can be met.
2) We have a statement of direction that the entire core team signed
on to. You can find it at http://www.ledgersmb.org/community/
part-way down the page.
3) Beyond the accounting bit, there are a lot of specific needs
individual businesses are going to have. These cannot be implemented
in a way that the application out of the box will work in every
possible vertical market, so instead, the infrastructure approach
allows the application to work in such an environment. Therefore I
think it is important to create an application where the community can
take it to any set of users regardless of their needs beyond
GAAP-compliant accounting.
>
> While each language has it's strengths and weaknesses, both Perl and Python
> are mature, fully functional, and have a whole host of toolkits available to
> them. Personally Python has a very slight advantage to me due to the syntax
> style, but that is mostly immaterial compared to user needs.
I have worked with both Perl and Python, have encountered
cross-platform issues in both, encountered things I really like about
both languages, and encountered things I really *don't* like about
each language. To me, as you say, the question is what the cost is of
switching languages, given where we are now and what people are using
the software.
While I don't think that the choice to switch languages is to be
lightly done, I can easily imagine a future where there are
implementations in different languages and that these compete (just as
we see with, say, Jabber).
>
> Have we taken a "holistic" look at LSMB and user needs? For example, would
> ecommerce integration be a desired need?
At least one core member is building e-commerce connectors for a customer.
> Or integration with a popular CMS?
Haven't heard a request for this yet.
> Or is enhanced light manufacturing a need being expressed?
Yes, this is a need.
> I suggest we
> first prioritize user needs and let them be the "basic" drivers in a
> discussion on the future of LSMB.
Currently the way that this is being done is that customers are
commissioning features and we work to help ensure that these features
are as forward-compatible as is reasonably possible. At the same
time, we are redesigning every aspect of the software, and are
engaging in detailed discussions with the community about needs in
specific modules.
Hope this helps,
Chris Travers